A fuzzy and neuro-fuzzy approach to modelling cattle grazing on pastures with low stocking rates in Central Europe

摘要: Abstract This paper describes a fuzzy and neuro-fuzzy approach to spatial explicit modelling of cattle grazing intensity in temperate zones Central Europe on pastures with low stocking rates. The aim was create simple model based datasets that could be collected as easily possible order predict the intensity. Large-scale moderate has been introduced many areas across cost management initiative for restoration or conservation open landscapes. As consequence heterogeneity large size under investigation, relationships between various investigated factors are often poorly understood data have high degree uncertainty. A selected which allowed operation vague knowledge Two rule-based models extensive presented this paper. first is Mamdani-type inference. linguistic rules were formulated by domain expert. second Sugeno-type given input–output dataset. number initial membership function parameters established effective clustering algorithm (subtractive clustering). Using neuronal network technique (ANFIS), optimal identified. These minimize root mean squared error model. dataset pasture North Germany used training an ANFIS learning procedure. In check generalization capability two models, predicted both compared gathered three over years. Mamdani- implemented using Fuzzy Logic Toolbox MATLAB©. results confirm suitability these grazing.